|
|
论文:2022,Vol:40,Issue(4):739-745 |
|
|
引用本文： |
|
|
陈禹乐, 李博, 梁红, 杨长生. 小样本下基于深度学习的声呐图像分类研究[J]. 西北工业大学学报 |
|
|
CHEN Yule, LI Bo, LIANG Hong, YANG Changsheng. Research on sonar image few-shot classification based on deep learning[J]. Northwestern polytechnical university |
|
|
|
|
|
|
|
| 小样本下基于深度学习的声呐图像分类研究 |
|
| 陈禹乐1, 李博2, 梁红1, 杨长生1 |
|
1. 西北工业大学 航海学院, 陕西 西安 710072;
2. 中国船舶集团有限公司第705研究所 水下信息与控制重点实验室, 陕西 西安 710077 |
| 摘要: |
| 水下环境复杂多样,使得声呐成像模糊难以人工提取特征,同时声呐图像不易获取,数量远少于光学图像,导致了小样本情况下声呐图像分类网络的训练过拟合现象明显,识别准确率低。基于所建立的声呐图像数据集进行预处理后,提出一种改进的带有类别偏好的标签平滑正则化方法,对训练数据的标签进行优化,减轻网络的自信程度,并基于迁移学习中微调的方法利用光学图像对网络参数进行预训练和冻结,融合以上方法构建了一种小样本下的分类网络模型。仿真实验结果表明,优化后的网络模型取得了最佳分类识别准确率,有效抑制了过拟合现象,能够在小样本下实现精确分类声呐图像。 |
| 关键词:
声呐图像分类
卷积神经网络
标签平滑
迁移学习
|
|
| Research on sonar image few-shot classification based on deep learning |
|
| CHEN Yule1, LI Bo2, LIANG Hong1, YANG Changsheng1 |
|
1. School of Marine Science and Technology, Northwestern Polytechnical University, Xi'an 710072, China;
2. Key Laboratory of Underwater Information and Control, the 705 Research Institute of CSSC, Xi'an 710077, China |
| Abstract: |
| The underwater environment is complex and diverse, which makes it difficult to evolve traditional methods such as manually extracting features from blurred images. What's more, sonar images are so hard to be obtained that their number is far less than optical images, this case usually is called as few-shot, which leads to over fitting and low recognition accuracy of networks for sonar image classification. Based on the established sonar image data set after image preprocessing, a sonar image few-shot classification method with multi strategy optimization fusion is proposed in this paper. It is an improved label smooth regularization method with category preferences that can optimize the labels of training data and reduce the self-confidence of the network. And then based on the fine-tuning method in migration learning, some parameters of pre-learned models from optical images domain are utilized to help improve the performance in the sonar images domain. Finally, all the above three optimization strategies are combined. The simulation experiments in this study conclude that the optimal recognition accuracy can increase to 96.94%, which proves the multi strategy fusion can effectively suppresses the overfitting phenomenon and accurately classifies sonar images in the case of few-shot. |
| Key words:
sonar image
convolution neural network
label smooth
transfer learning
|
|
| 收稿日期: 2021-11-05
修回日期:
|
| DOI: 10.1051/jnwpu/20224040739 |
| 基金项目: 国家自然科学基金(61971354)资助 |
| 通讯作者: 梁红(1969-),女,西北工业大学教授,主要从事水下目标探测研究。e-mail:lianghong@nwpu.edu.cn
Email：lianghong@nwpu.edu.cn |
| 作者简介: 陈禹乐(1998-),西北工业大学硕士研究生,主要从事水下目标识别研究。
|
|
| 相关功能 |
|
|
|
|
| 作者相关文章 |
|
| 陈禹乐 在本刊中的所有文章 |
| 李博 在本刊中的所有文章 |
| 梁红 在本刊中的所有文章 |
| 杨长生 在本刊中的所有文章 |
|
|
|
|
|
|
|
|
参考文献: |
|
|
[1] LOPERA Tellez. Underwater threat recognition: are automatic target classification algorithms going to replace expert human operators in the near future?[C]//OCEANS 2019-Marseille, Marseille, France, 2019: 1-4
[2] 康春玉, 夏志军. 水下无源声呐目标听觉域张量特征提取方法[J]. 声学学报, 2020, 45(6): 824-829 KANG Chunyu, XIA Zhijun. Tensor feature extraction of underwater passive sonar target based on auditory model[J]. Acta Acustica, 2020, 45(6): 824-829 (in Chinese)
[3] HINTON G E, SAlAKHUTDINOV R R. Reducing the dimensionality of data with neural networks[J]. Science, 2006, 313(5786): 504-507
[4] WILLIAMS D. Underwater target classification in synthetic aperture sonar imagery using deep convolutional neural networks[C]//International Conference on Pattern Recognition, 2017: 2497-2502
[5] 季文韬. 基于深度迁移学习的声呐图像识别方法研究[D]. 大连: 大连理工大学,2021 JI Wentao. Sonar image recognition method based on deep transfer learning[D]. Dalian: Dalian University of Technology, 2021 (in Chinese)
[6] WEI Z, YANG J, MIN S. A method of underwater acoustic signal classification based on deep neural network[C]//2018 5th International Conference on Information Science and Control Engineering, Zhengzhou, China, 2019: 46-50
[7] 王磊, 陈越超, 王青翠, 等. 基于卷积残差网络的水下主动目标回波图像分类方法研究[J]. 声学与电子工程, 2021(1): 1-4 WANG Lei, CHEN Yuechao, WANG Qincui, et al. Research on underwater active target echo image classification method based on convolution residual network[J]. Acoustics and Electronic Engineering, 2021(1): 1-4 (in Chinese)
[8] GV A, SUS A, VS A, et al. CNN algorithm for plant classification in deep learning[J]. Materials Today: Proceedings,2021,46(9): 3684-3689
[9] AKAY Metin, DU Yong, SERSHEN C L, et al. Deep learning classification of systemic sclerosis skin using the mobilenetV2 model[J]. IEEE Open Journal of Engineering in Medicine and Biology, 2021, 2: 104-110
[10] 金磊磊, 梁红, 杨长生. 积神经网络的水下目标声呐图像识别方法[J]. 西北工业大学学报, 2021, 39(2): 285-291 JIN Leilei, LIANG Hong, YANG Changsheng. Sonar image recognition of underwater target based on convolutional neural network[J]. Journal of Northwestern Polytechnical University, 2021, 39(2): 285-291 (in Chinese)
[11] LI Z ,QI B ,LI C. 3D sonar image reconstruction based on multilayered mesh search and triangular connection[C]//2018 10th International Conference on Intelligent Human-Machine Systems and Cybernetics, 2018: 60-63
[12] FEIFEI L, FERGUS R, PERONA P. One-shot learning of object categories[J]. IEEE Trans on Pattern Analysis and Machine Intelligence, 2006, 28(4): 594-561
[13] 巩文静, 田杰, 李宝奇, 等. 基于改进MobilenetV2网络的声光图像融合水下目标分类方法[J]. 应用声学, 2021, 41(3): 462-470 GONG Wenjing, TIAN Jie, LI Baoqi, et al. Acoustic-optical image fusion underwater target classification method based on improved MobilenetV2[J]. Journal of Applied Acoustics,2021, 41(3): 462-470 (in Chinese)
[14] 盛子旗, 霍冠英. 样本仿真结合迁移学习的声呐图像水雷检测[J]. 智能系统学报, 2021, 16(2): 385-392 SHENG Ziqi, HUO Guanying. Detection of underwater mine target in sidescan sonar image based on sample simulation and transfer learning[J]. CAAI Transactions on Intelligent Systems, 2021, 16(2): 385-392 (in Chinese)
[15] HE K, ZHANG X, REN S, et al. Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, 2016: 770-778
[16] RAFAEL Muller, SIMON Kornblith, GEOFFREY E Hinton. When does label smoothing help[C]//Advances in Neural Information Processing Systems, 2019: 4696-4705
[17] NIU S, LIU Y, WANG J, et al. A decade survey of transfer learning(2010-2020)[J]. IEEE Trans on Artificial Intelligence, 2020, 1(2): 151-166
[18] TAN C, SUN F, KONG T, et al. A survey on deep transfer learning[C]//International Conference on Artificial Neural Networks and Machine Learning, 2018: 270-279 |
|
|
|
相关文献: |
|
|
1．王耀力, 刘晓慧, 李斌, 常青.流形嵌入的选择性伪标记与小样本数据迁移[J]. 西北工业大学学报, 2021,39(5): 1122-1129 |
|
|
2．张飞, 闫杰.基于深度卷积神经网络的云分类算法[J]. 西北工业大学学报, 2020,38(4): 740-746 |
|
|
3．吕娜, 周家欣, 陈卓, 陈旿.小训练样本条件下的机载网络流量识别方法[J]. 西北工业大学学报, 2020,38(5): 1129-1137 |
|
|
|
|
|
|
